@moofoo For me a big part of growing towards "natural sound" was going for higher and higher efficiency speakers. The common view of high efficiency speakers is that they give you "more SPL for your watts". However, there's something fundamental happening: a 112dB/Wm loudspeaker gives us the same dynamic resolution (dynamics shading / range) as the live event does. The lower we get with loudspeaker efficiency, the more compressed the reproduction is, and as a consequence it looses "life" compared to the real instrument. A loudspeaker with 102dB/Wm is although considered ultra-efficient but it already imposes 10dB of dynamic compression to what is encoded on the recording, hence, it will already sound a little less live (=more compressed) than the instrument it reflects. 10dB compression in itself is already significant, as the dynamics has been compressed 10 times in power. Sadly, that is on the verge of what is achievable today with a loudspeaker cabinet the size of a healthy refrigerator, and it's about the limit what a normal person can squeeze into a large room.
However, looking at an average loudspeaker today of reasonably high sensitivity of 89dB/Wm, we get a whopping 23dB of dynamic compression, that is, we get 0,5% of the actual dynamic range! Even worse, most of the current high end loudspeakers roll today around 83dB/Wm (most of the sensitivity ratings are vastly inflated - when you see dB/2.83V/m & low impedance dip, that's the warning that the dB/Wm is often 5-6dB below the dB/Vm spec.) With 83dB/Wm there is a 29dB dynamic compression, that is, the information is coming at an almost x1000 energy compression. This has two consequences:
1. it will not sound like the real instrument - will have less life, radiancy, freedom in the sound. Even though all the textural information is there, it seems to be "resolved", and images sharply, but the dynamic information is severely compressed. Changing volume does not alter this fact, just changes the SPL at which this compressed material rides.
2. When set to same peak SPL as the real instrument, it will sound way more dense, forceful, "tight", "heavy", and tiring as to achieve the same peak SPL, the median energy output has to be much higher. The sounds that would have been soft are now compressed into the loud band. This makes sounds in the middle of the dynamic range more pronounced, and thereby the listener does experience it a an enhancement of sound quality by hearing fainter sounds more prominent. Yet, the very soft part of the dynamic range is lost, and the very low detail level is gone. This is obvious for experienced listeners, and is perhaps the most important aspect that distances real music instrument sound from recorded / played back compressed material.
Cheers,
Janos